The U.S. government has directly intervened in commercial AI availability, using export controls to force a global model shutdown just days after Anthropic filed for IPO. This sets a regulatory precedent that every AI vendor relationship now carries.
- Fable 5 and Mythos 5 disabled worldwide after a U.S. government export control directive arrived at 5:21 pm ET on June 12.
- Fable 5 was the most capable public AI model on the market according to Vals AI benchmarks, just three days after its commercial launch.
- Anthropic publicly disagrees with the recall, calling it disproportionate to a narrow jailbreak finding.
- Every user worldwide lost access because Anthropic has no technical way to filter its customer base by nationality.
- The shutdown landed as Anthropic’s confidential IPO process was underway, adding direct financial and investor risk to an already bitter standoff.
What Happened
The U.S. government ordered Anthropic on June 12 to immediately disable Fable 5 and Mythos 5 under national security authorities. Anthropic received the directive at 5:21 pm ET and complied within hours.
The order is framed as an export control action targeting foreign nationals inside or outside the United States. Because Anthropic has no way to filter users by nationality, access was cut for every customer worldwide.
Fable 5 had launched three days earlier and was ranked the most capable public AI model by Vals AI benchmarks. It was built as a safety-guardrailed version of Mythos, with capability restrictions covering high-risk areas like cybersecurity and biology.
Mythos had been shared with roughly 50 vetted organizations through Project Glasswing, including Amazon, Apple, Google, Microsoft, and CrowdStrike. Anthropic launched the restricted program after the model identified flaws in every major operating system and browser it tested.
Why It Matters
For SaaS teams that have built on Claude, this order introduces a supply availability risk that has no real parallel in commercial software. A vendor relationship with Anthropic now carries regulatory exposure that no contract or SLA clause can anticipate.
The skeptic case is blunt: Anthropic’s own public warnings about cybersecurity capability gave regulators precise technical language to justify the action. If you publicly state your model can break every major operating system, the government holds a basis to act on that statement.
We disagree that the finding of a narrow potential jailbreak should be cause for recalling a commercial model deployed to hundreds of millions of people.
Anthropic, Company Statement, Anthropic
Bottom Line
Watch Anthropic’s IPO process closely. The company filed a confidential S-1 with the SEC just days before the shutdown, and a prolonged government standoff will force hard answers from institutional investors before they commit any capital.
Any SaaS team relying on Anthropic’s frontier models should run a vendor continuity review this week. Track how this standoff develops through Relve, an AI trends intelligence platform built for founders watching moves like this one.
